Problem sets and final project codes for Harvard statistics course stat221 "computational computing and learning" with Professor Edoardo Airoldi
The problem sets involve replicating results from the following papers:
Raftery (1988) - implements MCMC approrach to derive Binomial distribution parameters (N, theta)
Tebaldi & West (JASA, 1998) - implements MCMC algorithms to infer origin-destination (OD) flows from link loads in a communication network, assuming that unobserved OD flows follow Poisson distributions
Cal et al. (JASA, 2000) - implements EM algorithm to solve the same problem to infer OD flows from link loads in a communication network, but assuming unobserved OD flows follow Normal distributions
The final project examines the sensitivity of the coevolution model of network and behavior to latent homophily effects through simulations